Brain-inspired semantic data augmentation for multi-style images
Data augmentation is an effective technique for automatically expanding training data in deep learning. Brain-inspired methods are approaches that draw inspiration from the functionality and structure of the human brain and apply these mechanisms and principles to artificial intelligence and compute...
| Prif Awduron: | , , |
|---|---|
| Fformat: | Erthygl |
| Iaith: | English |
| Cyhoeddwyd: |
Frontiers Media S.A.
2024-03-01
|
| Cyfres: | Frontiers in Neurorobotics |
| Pynciau: | |
| Mynediad Ar-lein: | https://www.frontiersin.org/articles/10.3389/fnbot.2024.1382406/full |
| _version_ | 1827308910431698944 |
|---|---|
| author | Wei Wang Zhaowei Shang Chengxing Li |
| author_facet | Wei Wang Zhaowei Shang Chengxing Li |
| author_sort | Wei Wang |
| collection | DOAJ |
| description | Data augmentation is an effective technique for automatically expanding training data in deep learning. Brain-inspired methods are approaches that draw inspiration from the functionality and structure of the human brain and apply these mechanisms and principles to artificial intelligence and computer science. When there is a large style difference between training data and testing data, common data augmentation methods cannot effectively enhance the generalization performance of the deep model. To solve this problem, we improve modeling Domain Shifts with Uncertainty (DSU) and propose a new brain-inspired computer vision image data augmentation method which consists of two key components, namely, using Robust statistics and controlling the Coefficient of variance for DSU (RCDSU) and Feature Data Augmentation (FeatureDA). RCDSU calculates feature statistics (mean and standard deviation) with robust statistics to weaken the influence of outliers, making the statistics close to the real values and improving the robustness of deep learning models. By controlling the coefficient of variance, RCDSU makes the feature statistics shift with semantic preservation and increases shift range. FeatureDA controls the coefficient of variance similarly to generate the augmented features with semantics unchanged and increase the coverage of augmented features. RCDSU and FeatureDA are proposed to perform style transfer and content transfer in the feature space, and improve the generalization ability of the model at the style and content level respectively. On Photo, Art Painting, Cartoon, and Sketch (PACS) multi-style classification task, RCDSU plus FeatureDA achieves competitive accuracy. After adding Gaussian noise to PACS dataset, RCDSU plus FeatureDA shows strong robustness against outliers. FeatureDA achieves excellent results on CIFAR-100 image classification task. RCDSU plus FeatureDA can be applied as a novel brain-inspired semantic data augmentation method with implicit robot automation which is suitable for datasets with large style differences between training and testing data. |
| first_indexed | 2024-04-24T19:20:01Z |
| format | Article |
| id | doaj.art-bc2b45d354834e4da44fab4106df9d12 |
| institution | Directory Open Access Journal |
| issn | 1662-5218 |
| language | English |
| last_indexed | 2024-04-24T19:20:01Z |
| publishDate | 2024-03-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neurorobotics |
| spelling | doaj.art-bc2b45d354834e4da44fab4106df9d122024-03-26T04:20:51ZengFrontiers Media S.A.Frontiers in Neurorobotics1662-52182024-03-011810.3389/fnbot.2024.13824061382406Brain-inspired semantic data augmentation for multi-style imagesWei WangZhaowei ShangChengxing LiData augmentation is an effective technique for automatically expanding training data in deep learning. Brain-inspired methods are approaches that draw inspiration from the functionality and structure of the human brain and apply these mechanisms and principles to artificial intelligence and computer science. When there is a large style difference between training data and testing data, common data augmentation methods cannot effectively enhance the generalization performance of the deep model. To solve this problem, we improve modeling Domain Shifts with Uncertainty (DSU) and propose a new brain-inspired computer vision image data augmentation method which consists of two key components, namely, using Robust statistics and controlling the Coefficient of variance for DSU (RCDSU) and Feature Data Augmentation (FeatureDA). RCDSU calculates feature statistics (mean and standard deviation) with robust statistics to weaken the influence of outliers, making the statistics close to the real values and improving the robustness of deep learning models. By controlling the coefficient of variance, RCDSU makes the feature statistics shift with semantic preservation and increases shift range. FeatureDA controls the coefficient of variance similarly to generate the augmented features with semantics unchanged and increase the coverage of augmented features. RCDSU and FeatureDA are proposed to perform style transfer and content transfer in the feature space, and improve the generalization ability of the model at the style and content level respectively. On Photo, Art Painting, Cartoon, and Sketch (PACS) multi-style classification task, RCDSU plus FeatureDA achieves competitive accuracy. After adding Gaussian noise to PACS dataset, RCDSU plus FeatureDA shows strong robustness against outliers. FeatureDA achieves excellent results on CIFAR-100 image classification task. RCDSU plus FeatureDA can be applied as a novel brain-inspired semantic data augmentation method with implicit robot automation which is suitable for datasets with large style differences between training and testing data.https://www.frontiersin.org/articles/10.3389/fnbot.2024.1382406/fulldata augmentationdeep learningrobust statisticsstyle transferuncertainty modelingbrain-inspired computer vision |
| spellingShingle | Wei Wang Zhaowei Shang Chengxing Li Brain-inspired semantic data augmentation for multi-style images Frontiers in Neurorobotics data augmentation deep learning robust statistics style transfer uncertainty modeling brain-inspired computer vision |
| title | Brain-inspired semantic data augmentation for multi-style images |
| title_full | Brain-inspired semantic data augmentation for multi-style images |
| title_fullStr | Brain-inspired semantic data augmentation for multi-style images |
| title_full_unstemmed | Brain-inspired semantic data augmentation for multi-style images |
| title_short | Brain-inspired semantic data augmentation for multi-style images |
| title_sort | brain inspired semantic data augmentation for multi style images |
| topic | data augmentation deep learning robust statistics style transfer uncertainty modeling brain-inspired computer vision |
| url | https://www.frontiersin.org/articles/10.3389/fnbot.2024.1382406/full |
| work_keys_str_mv | AT weiwang braininspiredsemanticdataaugmentationformultistyleimages AT zhaoweishang braininspiredsemanticdataaugmentationformultistyleimages AT chengxingli braininspiredsemanticdataaugmentationformultistyleimages |